The ‘small RNAs’ are generally referred to as small non-coding RNA molecules which are less than 300 nucleotides in length (Hagemann-Jensen et al., 2018). There are different classes of small RNAs which includes transfer RNAs (tRNAs), small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), miRNAs, PIWI-interacting RNAs (piRNAs), small interfering RNAs (siRNAs), transcription initiation RNAs (tiRNAs) and splice site RNAs (spliRNAs) (Morris and Mattick, 2014). These small RNAs participate in processes like RNA translation, RNA splicing, RNA modifications, mRNA destabilization or degradation, epigenetic processing, gene silencing, transcription initiation and splicing mechanisms (Hagemann-Jensen et al., 2018, Morris and Mattick, 2014). In short, they are involved in the regulation of the genome organization and gene expression. Therefore, the functional role of these small RNAs needs to be investigated further. In this thesis, we will discuss in details about the role of miRNAs in general and in relation to neuroblastoma.
18 MicroRNAs: Biogenesis and mode of action
MiRNAs are a large family of short (~22-25 nucleotides), endogenous, non-coding RNAs, which binds the partial or perfect complementary sequences in the 3’-untranslated region (UTR) of target messenger RNAs (mRNAs) leading to translational repression or mRNA degradation (Croce and Calin, 2005). Mounting evidence have established the roles of miRNAs in regulation of important cellular processes like survival, proliferation, metastasis, development, apoptosis and stress response (Croce and Calin, 2005). In 1993, the first miRNA was discovered while studying the development timing of the nematode Caenorhabditis elegans (Lee et al., 1993). Since then thousands of miRNAs have been identified across different species and the number is still increasing (Table 1).
Table 1: List of microRNA databases
The biogenesis of miRNAs takes place in a sequential manner which starts in the nucleus and ends in the cytoplasm (Figure 3). The miRNA-genes are mostly transcribed in the nucleus by RNA polymerase II (Pol II) enzyme into long primary miRNAs (pri-miRNAs) characterized by unique hairpin structure with 5’-cap and polyadenylated tail (Lee et al., 2004). The microprocessor complex (drosha ribonuclease III; DROSHA and its essential co-factor DiGeorge critical 8; DGCR8) further crops these pri-miRNAs into ~70-100 nucleotides long precursor miRNAs (pre-miRNAs) (Gregory et al., 2004, Denli et al., 2004). However, an alternative ‘splicing machinery’ has been reported for intronic miRNAs (called Mirtrons) which does not involve drosha-mediated cleavage (Ruby et al., 2007, Berezikov et al., 2007). Mirtrons have been discovered in several species including mammals, fruit-fly, Drosophila melanogaster and the nematode, Caenorhabditis elegans (Winter et al., 2009). After nuclear processing, the pre-miRNAs produced are then exported to the cytoplasm by exportin-5 (XPO5) in complex with GTP-binding nuclear protein, RAN (Ran-GTP) (Yi et al., 2003).
Database Link Reference
deepBase http://deepbase.sysu.edu.cn/ (Yang et al., 2010) miRGen2.0 http://www.microrna.gr/mirgen/ (Alexiou et al., 2010) miRNAMap http://miRNAMap.mbc.nctu.edu.tw/ (Hsu et al., 2006)
miRBASE http://microrna.sanger.ac.uk/ (Griffiths-Jones, 2006)
19 Figure 3: The microRNA biogenesis pathway
20
In the cytoplasm, an RNase III enzyme-DICER1 along with transactivation-responsive RNA-binding protein (TRBP) cleaves the pre-miRNA into an approximately 22 nucleotides long double-stranded (ds) miRNA with 2-nucleotide 3’ overhangs (Chendrimada et al., 2005). For some miRNAs, an additional endonuclease step by argonaute protein 2 (AGO2) cleaves the pre-miRNA generating the nicked Ago2-cleaved-precursor-miRNA (ac-pre-miRNA), which may facilitate the strand dissociation of mature miRNA (Diederichs and Haber, 2007).After DICER1 mediated cleavage, the ds miRNA is unwinded by helicases (like p68, p72, RNA helicase A, RCK/p54, TNRC6B, Gemin3/4 and Mov10) into single stranded mature miRNA (the guide strand) and the complementary passenger strand is subsequently degraded (Winter et al., 2009). The mature miRNA is then incorporated into the miRNA-induced silencing complex (miRISC) containing the argonaute proteins (AGO1, AGO2, AGO3 or AGO4) together with the members of GW182 family proteins and accessory factors. The mature miRNA-miRISC complex recognizes the complementary sequences in the 3’-UTRs of target mRNAs leading to mRNA degradation, destabilization or translational repression (Gregory et al., 2005, Winter et al., 2009).
Studies have, however, shown that the mature miRNA can also bind the 5’-UTR or the open reading frame (ORF) of the mRNA (Lytle et al., 2007, Moretti et al., 2010). In addition, instead of its usual function of guiding argonaute protein complexes for target mRNA silencing, miRNAs have been shown to act independently of argonaute proteins by interacting directly with ribonuceloproteins (decoy activity) (Beitzinger and Meister, 2010). Moreover, miRNAs have also been shown to directly interact with DNA and regulate the gene expression at transcriptional level (Kim et al., 2008).
Given the nature of miRNA and its interaction with target mRNA, it is not surprising that a single miRNA can target multiple genes. This regulatory function of miRNAs can thus affect many cellular pathways controlling important developmental and oncogenic processes.
Scientists have developed various different bioinformatic tools to predict miRNA targets (Table 2). Some of the predicted miRNA targets have been experimentally validated in various cancer types, which suggest a global role of miRNA regulation in cancer (Iorio and Croce, 2012).
21 Table 2: List of microRNA target prediction tools
Tool Link Reference
miRTarBase http://miRTarBase.mbc.nctu.edu.tw/ (Chou et al., 2018)
miRDB http://mirdb.org (Wong and Wang, 2015)
DIANA-microT http://www.microrna.gr/webServer (Paraskevopoulou et al., 2013)
miRWalk http://mirwalk.uni-hd.de/ (Dweep et al., 2011)
microRNA.org http://www.microrna.org (Betel et al., 2008) Targetscan http://www.targetscan.org (Lewis et al., 2005)
MicroRNAs and their roles in human cancer
In 2002, Calin and colleagues reported the first study identifying the involvement of miRNAs in cancer. They observed a frequent deletion and downregulation of chromosomal region 13q14 in B-cell chronic lymphocytic leukemia (CLL). This chromosomal region, which harbored miR-15 and miR-16 miRNA genes, was deleted or downregulated in about 68% of patients with CLL (Calin et al., 2002). The same group reported in 2004 that about 50% of miRNA genes are mapped to cancer-associated genomic regions or in fragile sites. They also demonstrated that the miRNAs located in the deleted regions have low levels of expression in many cancers (Calin et al., 2004).
Numerous studies have demonstrated miRNAs acting as oncogenes and/or tumor suppressors and affecting the different hallmarks of cancer (Hanahan and Weinberg, 2000) like sustaining proliferative signaling (Si et al., 2007, Medina et al., 2010), activating invasion and metastasis (Gregory et al., 2008, Chen et al., 2011), inducing angiogenesis (Hua et al., 2006) and resisting cell death (Lima et al., 2011) (Figure 4). Thus, the deregulated expression of miRNAs can have significant impact on the normal functioning of the cellular processes leading to the diseased condition or cancer.
22 Figure 4: MicroRNAs targeting hallmarks of cancer
Dysregulation of miRNA expression is a common feature of cancer. MiRNAs are shown to be over-expressed or under-expressed, and this aberrant expression has been associated with cancer phenotype (Deng et al., 2008). Whole-genome miRNA expression profiling has been used to detect the global expression of miRNAs in tumor specimens relative to normal tissues (Lu et al., 2005). In addition, miRNA profiling not only distinguish between cancerous and normal tumors but also between parental and resistant tumors. For instance, in our study, we employed the next generation sequencing technique to identify differentially expressed miRNAs from six different parental and resistant neuroblastoma cell lines isolated before and after chemotherapy treatment. We observed a downregulated expression of several miRNAs in the resistant cell lines, which is in concordance with the earlier studies demonstrating a general downregulation of miRNAs in cancer (Roth et al., 2016, Williams et al., 2017)
Several methods are developed for detecting miRNAs like bead-based miRNA profiling, miRNA microarrays, RT-qPCR, in situ hybridization techniques and the recent high-throughput sequencing. Moreover, loss and gain-of-function studies have been established to study the biology of miRNAs by overexpressing or silencing of particular miRNAs with synthetic mimics or antagomirs, respectively (Iorio and Croce, 2012). Overall, these studies have demonstrated the potential of miRNAs to be used as diagnostic and prognostic markers in cancer.
23
Due to their small size, miRNAs are more stable and resistant to degradation. In addition, extra-cellular miRNAs can be easily detected and extracted from body fluids such as blood (total blood, plasma or serum), exosomes, and even from urine, saliva and sputum (Weber et al., 2010). These so called circulating miRNAs are associated with various pathophysiological conditions and can thus be used as prognostic biomarkers for early diagnosis. For example, Lawrie et al. (2008), were the first to detect increased levels of tumor associated miRNAs (miR-155, miR-210 and miR-21) in serum of patients with diffuse large B-cell lymphoma and that increased levels of these miRNAs correlated with improved relapse-free survival (Lawrie et al., 2008).
MiRNAs can also response to specific therapies. In cholangiocarcinoma cell lines, targeted inhibition of miR-21 and miR-200b led to increased sensitivity to gemcitabine. This was the first study demonstrating the involvement of miRNAs in modulating drug resistance in cancer cells (Meng et al., 2006).
MicroRNAs as oncogenes and tumor suppressors
As mentioned earlier, dysregulation of miRNAs can affect one or several cellular processes including survival, proliferation, invasion, migration, metastasis, differentiation and apoptosis by acting as oncogenes or tumor suppressor genes (Babashah and Soleimani, 2011).
Cancer cells generally show the abundance of specific oncogenic miRNAs (also called oncomiRs) and the loss of tumor-suppressor miRNAs (Figure 5) (Table 3) (Esquela-Kerscher and Slack, 2006).
The oncomiRs repress the tumor suppressor genes, and/or genes that control cell differentiation or apoptosis (Esquela-Kerscher and Slack, 2006, Lu et al., 2005, Babashah and Soleimani, 2011). For instance, miR-155 is over-expressed and acts as an oncomiR by targeting SH2 domain-containing inositol 5-phosphatase 1 (SHIP1) in acute myeloid leukemia (Xue et al., 2014). In breast cancer, miR-21 was highly over-expressed compared to matched normal breast tissues. Thus, knockdown of miR-21 by anti-miR-21 oligonucleotides, suppressed cell growth in vitro and tumor growth in xenograft mouse model probably by indirect regulation of BCL2 expression (Si et al., 2007).
The tumor-suppressor miRNAs negatively regulate protein-coding oncogenes and or genes that inhibit cell differentiation or apoptosis (Esquela-Kerscher and Slack, 2006, Lu et al., 2005, Babashah and Soleimani, 2011). For example, the tumor suppressor, miR-34a have been shown to target MYCN (Wei et al., 2008) and E2F transcription factor 3 (E2F3) and induce
24
apoptosis in neuroblastoma (Welch et al., 2007). In chronic lymphocytic leukemia, miR-15a and miR-16-1 are deleted or downregulated, however over-expression of these miRNAs in leukemic cell line model negatively regulated the expression of anti-apoptotic BCL2 protein (Cimmino et al., 2005).
Figure 5: MicroRNA as oncogenes and tumor suppressors
Depending on the cellular context, miRNAs could function either as oncomiRs or tumor-suppressors. For instance, the polycistronic 17-92 cluster (includes 17-3p, miR-17-5p, miR-18a, miR-19a, miR-19b-1, miR-20a and miR-92a-1) located at the genomic locus 13q31, was not only over-expressed in tumor-cell lines but also amplified in tumors in diffuse large B-cell lymphoma. In addition, in vivo studies of miR-17-92 over-expression in transgenic mouse model of human B-cell lymphoma resulted in aggressive tumors, indicating the oncogenic role of miR-17-92 miRNA in cancer progression (Ota et al., 2004, He et al., 2005).
However, in another study, the c-Myc induced miR-17-92 cluster targeted and decreased the expression of the E2F1, involved in transition of G1-S phase of cell cycle progression, suggesting a tumor suppressor activity of this miRNA cluster (O'Donnell et al., 2005).
MiRNAs can function in complex regulatory circuits and feedback mechanisms. They are shown to work together in groups and co-operate to regulate oncogenes necessary for tumor progression. In human burkitt lymphomas, a group of miRNAs targeting the c-Myc oncogene
25
were silenced, which led to the over-expression of c-Myc and its targets involved in proliferation and survival. Interestingly, over-expression of Myc led to repression of some c-Myc targeting miRNAs, indicating a feedback mechanism in regulation of c-Myc expression (Bueno et al., 2011).
Table 3: Key microRNAs with oncogenic and tumor suppressor roles in neuroblastoma
miRNA Regulation mRNA targets Function Reference
miR-17-92 Up-regulated DKK3 Oncogenic (De Brouwer et al., 2012) miR-34a Down-regulated MAP3K9 Tumor suppressive (Tivnan et al., 2011)
miR-21 Up-regulated PTEN Oncogenic (Chen et al., 2012)
miR-376c Down-regulated CCND1 Tumor suppressive (Bhavsar et al., 2018)
miR-380 Up-regulated TP53 Oncogenic (Swarbrick et al., 2010)
miR-193b Down-regulated MCL1,CCND1,MYCN Tumor suppressive (Roth et al., 2018)
miR-15a Up-regulated RECK Oncogenic (Xin et al., 2013)
miR-323a Down-regulated STAT3 Tumor suppressive Manuscript I
The expression and function of oncomiRs can be increased or upregulated by multiple mechanisms including genomic amplifications, activating mutations, transcriptional activation and loss of epigenetic silencing. In contrast, loss of tumor-suppressor miRNAs can be due to genetic deletions, in-activating mutations, transcriptional repression and epigenetic silencing mechanisms (Lujambio and Lowe, 2012). Overall, the regulation of miRNA expression in cancer is very complex and therefore the mechanisms pertaining to the deregulation of miRNAs are discussed in the next section.
Mechanisms of microRNA deregulation in cancer
The deregulation or differential expression of miRNAs in cancer is undisputed. Not a single universal mechanism but a combination of several different mechanisms operate to modulate the expression profiles of individual or group of miRNAs in cancer setting (Deng et al., 2008). The mechanisms of miRNA deregulation can be broadly categorized into structural genetic variations, epigenetic modifications, transcriptional deregulation and defects in the miRNA biogenesis machinery (Figure 6) (Lin and Gregory, 2015, Deng et al., 2008).
26
The structural genetic variations include the DNA-copy number alterations (amplifications, deletions, and translocations) which are implicated in modulating the expression of miRNAs in cancers (Lujambio and Lowe, 2012, Deng et al., 2008). For instance, in chronic lymphocytic leukemia frequent deletions of chromosomal region 13q14 harboring the miRNAs miR-15 and miR-16 were observed in more than 50% of patients (Calin et al., 2002). In another study, an amplification of C13orf25 locus at 13q31-32 containing seven miRNA polycistronic cluster was reported in lymphoma patients (Ota et al., 2004, Tagawa and Seto, 2005).
In addition to genomic alterations, transcriptional regulators also play an important role in modulating the expression of miRNAs in cancer. For example, the activation of tumor suppressor gene, tumor protein P53 (TP53) led to the significant upregulation of 34-miRNAs and downregulation of 16-miRNAs in a genome-wide screen for TP53-regulated miRNAs in cancer. Among the deregulated miRNAs, miR-34 showed a marked upregulation, which is a well-known tumor suppressor shown to target genes, involved in promoting cell growth and proliferation. In the same study, other miRNAs like tumor suppressive, let-7a targeting the oncogenes rat sarcoma (RAS), high mobility group AT-Hook 2 (HMGA2) and miR-15a/16 targeting the BCL-2 were also identified (Tarasov et al., 2007). In another example, the transcriptional factor encoded by the proto-oncogene c-Myc, directly activates the expression of oncogenic miR-17-92 cluster (O'Donnell et al., 2005). Interestingly, c-Myc has been shown to repress a broader set of miRNA expression in mouse models of B cell lymphoma (Chang et al., 2008).
Defects in the miRNA biogenesis machinery can also affect the miRNA expression. In the first step of miRNA biogenesis, RNA polymerase II transcribes pri-miRNAs from miRNA genes (Lee et al., 2004), which has been shown to be deregulated in several cancers. As mentioned earlier, different genetic abnormalities like deletions, amplifications, and translocations can alter the expression of miRNA genes (Lin and Gregory, 2015, Deng et al., 2008).
27 Figure 6: The mechanisms of microRNA deregulation
28
In the second step of miRNA processing pathway, microprocessor complex, which contains DROSHA and DGCR8 enzymes, cleaves the pri-miRNA to generate pre-miRNA (Gregory et al., 2004, Denli et al., 2004). The microprocessor components DROSHA and DGCR8 are often dysregulated in cancer. For instance, a study by Lin and colleagues identified the downregulation of DROSHA in advanced stage neuroblastoma tumors, which correlated with the global downregulation of the miRNAs and poor clinical outcome (Lin et al., 2010).
However, in cervical squamous cell carcinoma, the upregulation of DROSHA was observed to link with altered miRNA expression (Muralidhar et al., 2011). In the third step, the pre-miRNAs are exported to cytoplasm by XPO5 via RAN-GTP (Yi et al., 2003). A study by Melo and colleagues demonstrated that, an inactivating mutation in XPO5 results in trapping of pre-miRNAs in the nucleus, impairing miRNA biogenesis machinery (Melo et al., 2010). Defects have also been observed in DICER and its essential co-factor TRBP, which are responsible for further processing of pre-miRNAs into ~22 nucleotides mature miRNAs (Chendrimada et al., 2005). For instance, DICER1 was shown downregulated in neuroblastoma and it correlated with global downregulation of miRNAs and poor clinical outcome (Lin et al., 2010). However, DICER was upregulated in metastatic prostate adenocarcinoma with global upregulation of miRNA expression and correlated with the increase in clinical stages (Chiosea et al., 2006).
Frameshift mutations were identified in TRBP2, which caused decreased protein expression in sporadic and hereditary colorectal carcinomas (Melo et al., 2009).
Epigenetic modifications like DNA methylations (mostly occurs in CpG sequences) and histone covalent modifications (acetylation, methylation and phosphorylation) alter the chromatin structure and regulate the pattern of gene expression (Portela and Esteller, 2010).
Weber and colleagues thoroughly analyzed miRNA genes and demonstrated that 50% of miRNA genes are associated with CpG islands (short DNA sequences located at the gene promoter). Moreover, the frequency of miRNA gene promotor methylation was higher as compared to protein coding genes (Weber et al., 2007). Studies have reported that CpG islands associated with miRNA gene promoter are frequently hypermethylated in cancer leading to epigenetic silencing of tumor suppressor miRNAs. These miRNAs can be re-activated by the treatment of chromatin-modifying drugs such as inhibitors of DNA methylation and/or histone deacetylase (Lujambio et al., 2008). For example, miR-127, which targets the proto-oncogene BCL6, is hypermethylated in human cancer cells. Upon treatment with chromatin modifying agents, a strong upregulation of miR-127 was observed (Saito et al., 2006). In another study, Parodi et al., (2016) analyzed the methylation status of a set of miRNAs in neuroblastoma cell
29
lines and identified a subset of hypermethylated and downregulated miRNAs (34 and miR-124) whose targets have important roles in fundamental cell processes like growth and apoptosis (Parodi et al., 2016). Similar to DNA hypermethylation, DNA hypomethylation have been shown to upregulate the expression of oncogenic miRNAs. The let-7a-3 gene was hypomethylated in lung adenocarcinomas, leading to high expression of let-7a-3 having oncogenic functions (Brueckner et al., 2007). In addition to DNA methylation, histone modifications could also modulate the miRNA expression profiles in cancer (Guil and Esteller, 2009).
MicroRNAs and drug resistance in human cancer
The role of miRNAs in mediating drug resistance in neuroblastoma is poorly understood. Very few studies have reported the direct involvement of miRNAs in modulating the drug resistance mechanisms in neuroblastoma (Table 4).
In one of our miRNA profiling studies, we employed deep sequencing technique to identify 34-downregulated and 8-upregulated miRNAs differentially expressed in neuroblastoma cell lines isolated from six patients at diagnosis and at relapse following intensive treatments (Roth et al., 2016). In another study, Ayers and colleagues identified differential expression of miRNAs in chemoresistant cell line models (SH-SY5Y and UKF-NB-3) of neuroblastoma (Ayers et al., 2015).
Chen and collaborators found miR-21 as the first miRNA associated with drug resistance in neuroblastoma. They observed the increased expression of miR-21 in cisplatin-resistant SH-SY5Y and BE(2)-M17 neuroblastoma cells as compared to parental cells. Therefore, by using antagomir against miR-21, they knocked down the expression of miR-21, which sensitized the cisplatin-resistant cells. Further, they ectopically expressed pre-miR-21 in parental cells, which led to increased resistance to cisplatin treatment and enhanced proliferation by modulating the phosphatase and tensin homolog (PTEN) protein levels (Chen et al., 2012).
In another study, miR-204 was shown to increase sensitivity of neuroblastoma cell lines to cisplatin and etoposide. Surprisingly, miR-204 had no effect on neuroblastoma cell growth in the absence of chemotherapeutic agents. The miR-204 has been shown to directly target the 3’UTR sequence of BCL2 and an oncogene, neurotrophic receptor tyrosine kinase 2 (NTRK2) both of which are important mediators in facilitating resistance to several chemotherapeutic agents. Moreover, BCL2 and NTRK2 are significantly associated with poor patient survival in chemo-resistant neuroblastoma (Ryan et al., 2012).
30
Table 4: MiRNAs involved in modulating drug resistance in neuroblastoma
microRNA Drug/s Target/s Reference
miR-17-5p-92 Not determined p21 and BIM (Fontana et al., 2008)
miR-21 Cisplatin PTEN (Chen et al., 2012)
miR-204 Cisplatin, Etoposide BCL2 and NTRK2 (Ryan et al., 2012)
miR-137 Doxorubicin CAR (Takwi et al., 2014)
miR-520f Cisplatin, Etoposide NAIP (Harvey et al., 2015)
miR-155 Cisplatin TERF1 (Challagundla et al., 2015)
miR-497 Not determined CHEK1, AKT and VEGFA (Soriano et al., 2016)
miR-141 Cisplatin FUS (Wang et al., 2016)
miR-137 Doxorubicin CAR (Zhao et al., 2017)
Two independent reports highlighted the importance of miR-137 in modulating the doxorubicin sensitivity of neuroblastoma cells. Takwi and colleagues observed downregulated expression of miR-137 and an inverse high expression of constitutive androstane receptor (CAR) and MDR1 in doxorubicin-resistant neuroblastoma cells as compared to parental cells.
Furthermore, miR-137 was shown to directly target CAR and that over-expression of miR-137 led to sensitization of resistant cells to doxorubicin and reduction of the resistant tumor growth in vivo (Takwi et al., 2014). Zhao et al., (2017) proposed yet another mechanism to increase the sensitivity of neuroblastoma cells to doxorubicin. In this study, short-interfering RNA (siRNA) knockdown of histone deacetylase 8 (HDAC8) which is often upregulated and correlated with advance stage disease, increased sensitivity of neuroblastoma cells to doxorubicin via upregulation of miR-137 and inhibition of the MDR1 (Zhao et al., 2017).
Tumor microenvironment has been shown to play important role in mediating drug
Tumor microenvironment has been shown to play important role in mediating drug